From owner-freebsd-hackers Sun Mar 31 17:38:07 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id RAA16617 for hackers-outgoing; Sun, 31 Mar 1996 17:38:07 -0800 (PST) Received: from palmer.demon.co.uk (palmer.demon.co.uk [158.152.50.150]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id RAA16606 Sun, 31 Mar 1996 17:38:01 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by palmer.demon.co.uk (8.7.5/8.6.11) with SMTP id CAA00521 ; Mon, 1 Apr 1996 02:37:47 +0100 (BST) To: Sujal Patel cc: hackers@FreeBSD.ORG, macgyver@infinet.com, DARREND@novell.com, "Jordan K. Hubbard" , dlacroix@cray-ymp.acm.stuorg.vt.edu From: "Gary Palmer" Subject: Re: Netscape 3.0b2 [Solution & Summary] In-reply-to: Your message of "Sun, 31 Mar 1996 15:32:02 CDT." Date: Mon, 01 Apr 1996 02:37:47 +0100 Message-ID: <519.828322667@palmer.demon.co.uk> Sender: owner-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk Sujal Patel wrote in message ID : > > Alright folks... After getting a working trace of the way netscape was > running, I've figured out that they are sending an illegal request to our > Xserver. This causes netscapes forked Java process to exit aburptly. Why > this only happens to some of us is still a mystery, but the solution is > simple: Strange. I just ran ktrace on a netscape process (I did: ktrace -d -i netscape -install http://java.sun.com/ ) And would you believe it, this is the only sub-process I can find: 471 netscape.bin RET fork 470/0x1d6 471 netscape.bin CALL dup2(0xd,0x1) 471 netscape.bin RET dup2 1 471 netscape.bin CALL dup2(0xd,0x2) 471 netscape.bin RET dup2 2 471 netscape.bin CALL close(0) 471 netscape.bin RET close 0 471 netscape.bin CALL getdtablesize 471 netscape.bin RET getdtablesize 128/0x80 471 netscape.bin CALL close(0x7f) 471 netscape.bin RET close -1 errno 9 Bad file descriptor 471 netscape.bin CALL close(0x7e) 471 netscape.bin RET close -1 errno 9 Bad file descriptor [ You get the idea ... trimmed ] 471 netscape.bin CALL close(0xf) 471 netscape.bin RET close -1 errno 9 Bad file descriptor 471 netscape.bin CALL close(0xe) 471 netscape.bin RET close -1 errno 9 Bad file descriptor 471 netscape.bin CALL close(0xd) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0xc) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0xb) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0xa) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0x9) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0x8) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0x7) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0x6) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0x5) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0x4) 471 netscape.bin RET close 0 471 netscape.bin CALL close(0x3) 471 netscape.bin RET close 0 471 netscape.bin CALL execve(0xefbfbf2c,0xefbfc34c,0x49db00) 471 netscape.bin NAMI "/bin/netstat" 471 netscape.bin RET execve -1 errno 2 No such file or directory 471 netscape.bin CALL execve(0xefbfbf2c,0xefbfc34c,0x49db00) 471 netscape.bin NAMI "/usr/bin/netstat" The main process then goes on to do (some time later) : 470 netscape.bin CALL read(0xc,0x4ea000,0x1000) 470 netscape.bin GIO fd 12 read 836 bytes "Name Mtu Network Address Ipkts Ierrs Opkts Oerrs \ Coll ed0 1500 00.00.c0.5c.46.8e 0 0 1 0 \ 0 ed0 1500 10 10.0.0.1 0 0 1 0 \ 0 lo0 16384 372 0 372 0 \ 0 lo0 16384 127 127.0.0.1 372 0 372 0 \ 0 sl0* 1500 0 0 0 0 \ 0 sl0* 1500 10.0.1.1 10.0.1.1 0 0 0 0 \ 0 sl1* 552 0 0 0 0 \ 0 tun0 1524 883 0 1391 0 \ 0 tun0 1524 158.152 158.152.50.150 883 0 1391 0 \ 0 tun1* 1500 0 0 0 0 \ 0 " I have *NO* idea why it runs ``netstat -i'' - I doubt it can parse the output somehow! That (apart from /bin/sh initially running /usr/local/lib/netscape/netscape.bin) is the ONLY fork in the entire ktrace output... Actually, I wonder if my ktrace is broken somehow: 470 netscape.bin CALL write(0x2,0xefbf98c8,0x45) 470 netscape.bin GIO fd 2 wrote 69 bytes "Warning: Cannot allocate colormap entry for default background " Then later on: 470 netscape.bin CALL read(0xd,0x42ec94,0x3fe) 470 netscape.bin GIO fd 13 read 69 bytes "Warning: Cannot allocate colormap entry for default background " Process 470 wrote to fd 2, and then read it back in from fd 13?!? I'm going to look at this more closely. Gary P.S. At no time have I gotten this version of netscape to produce a ``java console''...